Effect of reference template selection in diffusion tensor based voxel-wise analysis

Introduction Diffusion tensor based voxel-wise analysis (DT-VBA) has been increasingly used to investigate subtle differences in brain white matter microstructure among various types of patient groups and groups of control subjects.[1-2] The method provides an inference of inter-group differences in white matter integrity on a voxel-wise basis in a standardized space. Spatial normalization is an important preprocessing step in DT-VBA analysis and is used to reduce inter-subject anatomical variability by registering each image volume to a standard space.[3-5] Varieties of reference template and processing procedures exist in the implementation of DT-VBA.[6] It is still not clear which image processing procedure may produce more plausible or more accurate results. The purpose of this study is to systematically evaluate the effects and accuracy of varying reference template and image preprocessing procedure of the DT-VBA.